Is the Proton a good reason to buy Life Technologies stock?

[steinmann]

Have been fairly impressed with their fast and steady progress on the PGM. Thats what makes me confident that they will continue to deliver when they release the Proton. So what are the relevant questions we need to ask?

Will they be able to outcompete Illumina?
Cost of the machine, speed and scalability make that faily likely. They are already beating them in terms of units sold with the PGM, a machine that looks pretty pale comared to the Proton. Recent price increases by Illumina seem to imply that their margins are already pretty thin.

Are there other relevant competitors?
Oxford Nanopore would be the only one that comes to my mind, but they have still to proove that their machine is more than vaporware.

Is the price right?
A P/E of 20 is certainly not cheap, but compared to Illuminas P/E of 72 a bargain.

Anything I missed? Different opinions? Did anyone else buy?
Bought for a good fraction of my net worth in the beginning of June.

[TonyBrooks]

At the moment we still use Illumina over PGM where possible. Around 50% of our costs are labour costs. If you compare the hands on time it takes to do a PGM run (ePCR, OneTouch enrichment, initialise instrument, load chip ~1 day) to set up our MiSeq ~15 minutes then it's a no brainer.
For large projects, Illumina again is the way to go as we can run those on a HiSeq or GA. For example, an RNA-seq experiment just isn't feasible on the PGM as the number of reads is far too low. Although, admittedly the Proton may change that.
The Achilles Heel with Illumina is still the diversity problem, so some projects (amplicon) are still more suited to PGM.

[krobison]

The short answer is no, simply because it is a mistake to buy any stock to such a degree and based on such information. The market is largely efficient, and much of what moves markets is not predictable (today being the 11th anniversary of a horrific exemplar of that statement).

Another thread on Proton had information with regard to their launch and some insanely bullish projections by a stock analyst. Ask yourself: to what degree am I just as delusional as this character.

By delusional, I refer to his crude extrapolation that Proton will enjoy the same success as HiSeq; maybe it will but probably it won't. Ion still lacks wide acceptance in academia, and that's still a huge market. Ion may have placed more units than MiSeq, but much murkier is whether the consumable sales (which is where the real money is) will follow. Numerous competitors (e.g. Oxford Nanopore, Genia & GnuBio) are waiting in the wings & may bite off pieces of the pie. It isn't clear how big the pie is.

You'll rest much easier at night if you take A Random Walk Down Wall Street. You won't get filthy rich, but you also won't find yourself dirt poor.

[steinmann]

Thanks a lot for your insightful comments!

Originally Posted by krobison

The short answer is no, simply because it is a mistake to buy any stock to such a degree and based on such information. The market is largely efficient, and much of what moves markets is not predictable (today being the 11th anniversary of a horrific exemplar of that statement).

Largely efficient, yes. But especially when it comes technology stock Mr. Market does all kinds of crazy things. I am working under the assumption that we who run and analyze data from these machines should be better judges on what technology prevails than the guys at wall street. Dont you think so?

Another thread on Proton had information with regard to their launch and some insanely bullish projections by a stock analyst. Ask yourself: to what degree am I just as delusional as this character.

Really good point! Actually the one I have been most concerned with for quite a while. Quite possible that as a technology geek I am somewhat blinded by the sexyness of their technology and excellent marketing

By delusional, I refer to his crude extrapolation that Proton will enjoy the same success as HiSeq; maybe it will but probably it won't. Ion still lacks wide acceptance in academia, and that's still a huge market. Ion may have placed more units than MiSeq, but much murkier is whether the consumable sales (which is where the real money is) will follow. Numerous competitors (e.g. Oxford Nanopore, Genia & GnuBio) are waiting in the wings & may bite off pieces of the pie.

All valid concerns, but at the end of the day its just about what technology offers the best bang for the buck. Preferably with short run times. Right now thats a race between Illumina and Life with Life clearly behind but accelerating at an insane pace. Think its highly unrealistic that any of the other competitors will just pop out of nowhere and offer something that can compete with those two. Only news from Oxford Nanopore since AGBT is that its investors agreed on another round of funding..

It isn't clear how big the pie is.

Somewhat bigger than Illuminas share today. If Illumina loses marketleadership to Life thats big enough.

You'll rest much easier at night if you take A Random Walk Down Wall Street. You won't get filthy rich, but you also won't find yourself dirt poor.

That would probably the best advice if I had a family to feed or where at an age where retirement planing was more relevant

[steinmann]

Originally Posted by TonyBrooks

At the moment we still use Illumina over PGM where possible. Around 50% of our costs are labour costs. If you compare the hands on time it takes to do a PGM run (ePCR, OneTouch enrichment, initialise instrument, load chip ~1 day) to set up our MiSeq ~15 minutes then it's a no brainer.

Yes, EM PCR is inferior to automated cluster generation. For a fair comparison though one would have to look at hands on time from quantified library to start of run. How much is that in your experience?

[Wallysb01]

Originally Posted by TonyBrooks

At the moment we still use Illumina over PGM where possible. Around 50% of our costs are labour costs. If you compare the hands on time it takes to do a PGM run (ePCR, OneTouch enrichment, initialise instrument, load chip ~1 day) to set up our MiSeq ~15 minutes then it's a no brainer.

~15 min??....um I don't think that's equivalent there.

[ymc]

LIFE is not a pure sequencing play like ILMN. Therefore I don't think it is fair to compare their P/Es.

Also, if you are a tech geek, you should know that the current PGM's read error rate is higher than MiSeq. PGMs also have homopolymer issues. The Proton that is supposedly going to come out this month has a presumed throughput as MiSeq. If it can improve its error rate to the level of MiSeq, I can then start to get excited about it.

[steinmann]

Originally Posted by ymc

LIFE is not a pure sequencing play like ILMN. Therefore I don't think it is fair to compare their P/Es.

Yes, but they should still be in the same ballpark which they are clearly not.

Also, if you are a tech geek, you should know that the current PGM's read error rate is higher than MiSeq. PGMs also have homopolymer issues.

True, but recently released datasets show that this is not much of a problem any more.

The Proton that is supposedly going to come out this month has a presumed throughput as MiSeq. If it can improve its error rate to the level of MiSeq, I can then start to get excited about it.

60-80 million reads for the Proton is substantially more than the 15-17 for the MySeq. Error rates for the new chips are of course still a big unknown. Hard to say how smaller beads and a therefore weaker signal will impact data quality.

[low99]

How does the Proton produce 60-80 million reads? My numbers don't add up.
The way I calculate reads is total output divided by bp.
So 10Gb/200bp = 50 million
Where are they getting the extra 30 million? Am I doing something wrong?

Can you do Pair Ends on Proton? Nothing has been said on that

Am I the only one to be disappointed in the fact that there was no Proton data presented at Ion World?

[steinmann]

The chip has 165 wells. Given a yield of 36-48% 60-80 million reads is what you should expect. I would guess that thats quite realistic and the 10 Gb spec is rather conservative.

I agree that its really disappointing that they havent uploaded a single dataset so far. Expected them to do that before they ship the instrument.

[adaptivegenome]

Originally Posted by steinmann

Yes, but they should still be in the same ballpark which they are clearly not.


True, but recently released datasets show that this is not much of a problem any more.

Which dataset is this? I have seen no data to suggest this, aside from hype in the popular press...

[steinmann]

This:
http://www.iontorrent.com/lomanpaper

And the recent 400 bp run on the 318 chip:
http://ioncommunity.lifetechnologies...tasets_by_chip

Indicate that they have been making fairly impressive improvements when it comes to error rates.

[adaptivegenome]

Originally Posted by steinmann

This:
http://www.iontorrent.com/lomanpaper

And the recent 400 bp run on the 318 chip:
http://ioncommunity.lifetechnologies...tasets_by_chip

Indicate that they have been making fairly impressive improvements when it comes to error rates.

I can't access the dataset without registration but we do have a PGM that we purchased late last year and the homopolymer errors (as well as other indel related errors) are more frequent than with Illumina, something confirmed by the paper you posted.

It will be exciting if this error is reduced on the new machines and particularly the Protons.

[steinmann]

Their claim is that the improvements were made through changes in chemistry (since introduction of the 300 bp kit) and base calling algorithm (torrent suite 3.0).

[Wallysb01]

Originally Posted by steinmann

Their claim is that the improvements were made through changes in chemistry (since introduction of the 300 bp kit) and base calling algorithm (torrent suite 3.0).

Have we seen any recent data verifying what Ion Torrent put out themselves? I know the rep that comes by our campus pushed this pretty hard, essentially claiming something like, "yes, we aren't as good as illumina with homopolymer issue, but we're good enough that it isn't a big problem anymore."

He showed some slide or another, but don't remember be hugely impressed. If I remember correctly, they still had problems with homopolymers out to 4 bp, though it was reduced from previously. And the situation got bad at 6bp.

[steinmann]

Yes, nothing but a claim. But a member of this forum suggested that he had indeed seen such improvements. Posted yesterday in this thread at 2:40 PM but post got lost. Maybe linked to recent instability if this site?

MrGuy has just replied to a thread you have subscribed to entitled - Is the Proton a good reason to buy Life Technologies stock? - in the Ion Torrent forum of SEQanswers.

This thread is located at:
http://seqanswers.com/forums/showthr...5&goto=newpost

Here is the message that has just been posted:
***************
From the data I've produced in the past month or two with the latest chemistries, I have no issues with the sequencing chemistry or specifically the homopolymer issue you are refering to.
***************

[nschuch]

Originally Posted by TonyBrooks

At the moment we still use Illumina over PGM where possible. Around 50% of our costs are labour costs. If you compare the hands on time it takes to do a PGM run (ePCR, OneTouch enrichment, initialise instrument, load chip ~1 day) to set up our MiSeq ~15 minutes then it's a no brainer.

Having prepared samples for Ion Torrent and Miseq I don't think this is a fully fair comparison. Washing and initializing the machines takes the same amount of time. Running ePCR and enrichment with the Ion Torrent is time consuming, but very easy. There's an hour of actual work involved going from prepared libraries to putting the chip in and running the Ion Torrent. Then there are five hours of waiting. You can work on other things during that time.

Paired end is another story. MiSeq is much easier to do paired end on. Considering that our Ion Torrent sales rep told me, in person, that they were disappointed by the quality of the second reads, I'm tempted not to even consider paired-end.

[MrGuy]

Originally Posted by steinmann

Yes, nothing but a claim. But a member of this forum suggested that he had indeed seen such improvements. Posted yesterday in this thread at 2:40 PM but post got lost. Maybe linked to recent instability if this site?

I deleted it since it was nothing but a claim. I'd rather support it with data for you, but didn't feel like making a big deal out of it. Regardless, works fine for what I do, no issues with the latest chemistries or protocols with regards to homopolymers.

[Mirth]

Hi All,

Very interesting thread. I came across this youtube video that shows how Ion Torrent has improved with respect to accuracy.

http://www.youtube.com/watch?v=g2s-1...feature=relmfu

[Mirth]

More on accuracy check out time point 18:29

http://www.youtube.com/watch?v=HJ-UX...feature=relmfu